If you've used Yahoo Finance or any other free stock screener, you know the routine: pick 3-5 ratio thresholds (P/E < 15, dividend yield > 3%, debt-to-equity < 1, ROE > 15%), click "screen", and get a list of 50 stocks that pass all five.
The list is usually 60-70% value traps.
Why? Because a generic screener treats ratios as independent filters. Real value investing is a multi-step decision tree where each step depends on the previous step's answer. Buffett doesn't apply 5 filters in parallel - he applies 5 questions in a specific order, and the answer to question 3 changes what counts as a pass on question 4.
This post explains the difference, with concrete examples of stocks where a generic screener and the Buffett framework disagree.
What a generic screener does
A typical "value stock" screen on Yahoo Finance or Finviz applies independent filters:
- P/E < 15 ✓
- P/B < 1.5 ✓
- Current ratio > 2 ✓
- Debt/equity < 0.5 ✓
- Dividend yield > 3% ✓
A stock that passes all five gets surfaced. The screen has no concept of quality, moat, capital allocation, or durability. It just checks five boxes.
The trouble: there are roughly 50-80 stocks in any given quarter that pass all five filters. Of those:
- ~40% are cyclicals at the cycle top (cheap by trailing P/E, expensive by normalised)
- ~20% are dying businesses with declining revenue (cheap because the market sees the decline coming)
- ~15% are leveraged businesses where the debt/equity ratio looks fine because of accounting tricks (operating leases, off-balance-sheet vehicles)
- ~25% are genuine value opportunities
The screener cannot distinguish between these four categories. To a screener, all 80 names look identical.
What the Buffett framework does
The Buffett framework, properly applied, is a decision tree, not a filter:
Step 1: Can I understand this business? (qualitative gate, not in any screen)
If no → reject. If yes → continue.
Step 2: Does it have a durable moat?
Check sustained ROIC over 10 years. If declining trend → reject (eroding moat). If stable or improving → continue.
Step 3: Will this business be valuable in 10 years?
Check gross-margin stability, customer concentration, technological-disruption risk. If margin trending down or one-customer-dependence → reject.
Step 4: Is management trustworthy?
Check capital-allocation history, buyback timing, M&A track record. If destructive → reject.
Step 5: Is the financial structure sound?
Check current ratio, debt levels in context of business cyclicality. If fragile → reject.
Step 6: Now - is the price reasonable?
Check owner-earnings yield. This is the LAST step. If yes → buy. If no → wait.
A screener that runs P/E < 15 as the FIRST filter is asking the wrong question first. Price is the last question, not the first.
Worked example: AT&T (T) vs Costco (COST)
In 2022, both stocks pass a basic "high yield + reasonable P/E" screen:
- AT&T: P/E 8, dividend yield 7%, debt/equity 1.0, ROE 12%
- Costco: P/E 38, dividend yield 0.5%, debt/equity 0.4, ROE 25%
A generic screener would surface AT&T and reject Costco. AT&T looks cheap; Costco looks expensive.
Apply the Buffett framework decision tree:
AT&T:
- Step 2 (moat): declining moat - telecom is a regulated commodity with little pricing power
- Step 3 (durability): aggressive declining (cord-cutting, 5G overinvestment)
- Step 4 (management): mediocre capital allocation history (WarnerMedia + Time Warner + DirecTV value destruction)
- → Rejected at step 2 or 3. Price never enters the conversation.
Costco:
- Step 2 (moat): wide and strengthening - membership-based, scale economics
- Step 3 (durability): high - staples consumption + Kirkland brand strength + member-renewal loyalty
- Step 4 (management): excellent - disciplined buyback timing, no destructive M&A, Jim Sinegal legacy
- Step 5 (health): strong - fortress balance sheet
- Step 6 (price): high but justifiable for the durability. Partial pass.
- → Accepted with valuation caveat (wait for pullback, or accept rich multiple).
In 2026, AT&T has paid out its 7% yield for 4 years while losing 30%+ of the principal value. Costco has compounded at ~15%/year over the same period. The screener said AT&T; the framework said Costco. The framework was right.
Worked example #2: Kraft Heinz (KHC) vs Coca-Cola (KO)
Both score similarly on a basic screen:
- KHC: P/E 14, yield 4.5%, debt/equity 0.6
- KO: P/E 27, yield 3.0%, debt/equity 0.7
Generic screener: KHC looks cheaper.
Buffett framework:
KHC:
- Step 2 (moat): eroding - private-label substitution accelerated 2017-onwards
- Step 3 (durability): declining - gross margin trend down 200 bps over 5 years
- → Rejected at step 2/3, regardless of P/E
KO:
- Step 2: durable - global distribution moat is one of the widest in consumer goods
- Step 3: extreme - Coca-Cola has been gradually expanding share for 100+ years
- Step 5: solid
- Step 6: rich P/E but justified by extreme durability
KO has 6/7 framework consensus (Lynch fails on growth rate). KHC has 3/7 (Smith and Fisher fail on margin compression). Generic screener says both pass; framework says only one does.
Why invest-like exists in this exact space
Generic screeners are tools for finding candidates. They surface 50-80 names that pass independent filter criteria. They cannot evaluate which of those names are genuine value.
invest-like is built on the opposite assumption: the right output isn't a list, it's a verdict per stock that integrates the decision tree above. Our 5-pillar Buffett Brain verdict applies the steps in order; our 7-framework consensus screens against multiple decision trees simultaneously; the Boardroom shows you what happens when the trees disagree.
The cohort backtest shows the result: stocks passing the 7-of-7 consensus (the strict decision-tree filter) returned a median +73.6% above the S&P 500 over 5 years, with 81% of the cohort beating the index. Stocks that pass a generic screener would have returned much closer to the index, because the screener can't distinguish KO from KHC.
When to use each tool
Use a generic screener when:
- You want to discover candidate names you haven't heard of
- You want to filter the universe to a manageable list before deeper analysis
- You're testing a specific hypothesis (e.g., "find me all stocks with P/B < 1 in the energy sector")
Use the Buffett framework / invest-like when:
- You have a candidate name and want a verdict on it
- You want to verify the consensus across multiple frameworks
- You want to know the specific reason a framework rejects a name
- You want to see the Boardroom debate on a stock you're considering
The two are complementary. A common workflow: screen on Finviz to find candidates → drop each into /buffett/[ticker]/ to get the verdict → open the Boardroom on the highest-conviction picks.
Disclosure
Educational tool. AT&T, Coca-Cola, Costco, Kraft Heinz are mentioned as analytical examples. Not recommendations. Past framework verdicts do not guarantee future stock performance. invest-like provides analysis, not personalised financial advice.
Author: Zaid Ghazal, founder of invest-like, indie SaaS, Kiel, Germany.